TOWARDS AN INDEX OF IDIOLECTAL SIMILITUDE 505
different genres, IIS should also be close to 1 (but perhaps not
as close as results in hypothesis 2).
D. MethodologyThe analysis of idiolectal distance that is presented here is
based on research carried out in two projects.^25 Each project
involved several stages where different numbers of subjects and
methods were analyzed. This article presents final results
obtained in the last stage, in which six individuals were studied
per each module, and a final list of variables, ranging between
10 and 18 depending on the module, were selected after some
preliminary studies where some other variables were discarded.
Moreover, a total of four different methods were explored, but
only three were involved in the final stage. The remaining
method, which was based on the Euclidean distance, was finally
discarded, and it is not included in this account.
- Linguistic Modules and Variables
The protocol devised to calculate the IIS has explored, so
far, three different linguistic levels, or modules: the
phonological module, the morphosyntactic module, and the
discourse-pragmatic module. The phonological module involves
the analysis of phonological processes related to insertion,
elision, or change of sounds, such as yod-coalescence in English
(a process by which a word like duke can be pronounced [dju:k]
or [dʒu:k]). The morphosyntactic module considers variables
related to morphological and syntactic patterns, such as the
presence or absence of the conjunction that in a sentence like I
thought (that) it was nice. Finally, the discourse-pragmatic
module considers discursive and pragmatic phenomena, such as
the choice of the intensifier really in contrast with other
intensifiers such as absolutely or completely, as in I was
really/absolutely/completely terrified.
(^25) See supra note 6.